"Much better alternative would be that Proton makes it's own Algorithm."
This is a very bad idea. Encryption algorithms are security mechanisms, and the design of a security mechanism should be open rather than secret. Therefore, encryption algorithms should not be proprietary. Cryptographic libraries take years of dedicated testing and research to be considered safe, and new issues are constantly discovered.
"liboqs is integrated in TLS, means no additional work is required for the client side"
"Much better alternative would be that Proton makes it's own Algorithm."
This is a very bad idea. Encryption algorithms are security mechanisms, and the design of a security mechanism should be open rather than secret. Therefore, encryption algorithms should not be proprietary. Cryptographic libraries take years of dedicated testing and research to be considered safe, and new issues are constantly discovered.
"liboqs is integrated in TLS, means no additional work is required for the client side"
The OpenSSL 1.1.1 fork using liboqs is currently for experimental purposes only. It doesn't make a ton of sense to rush into this. Signal and iMessage are implementing Kyber so we'll have good data in a bit to ensure nothing breaks. Additionally, issues have already been discovered in Kyber, both in its math (https://blog.cr.yp.to/20231003-countcorrectly.html) and implementation (https://techmonitor.ai/hardware/quantum/kyberslash-kyber-vulnerabilities).